[SERVER-15850] DeadlockExceptions thrown during multi-writes may trigger stale configs Created: 28/Oct/14  Updated: 10/Dec/14  Resolved: 04/Nov/14

Status: Closed
Project: Core Server
Component/s: Concurrency, Sharding
Affects Version/s: 2.7.8
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Greg Studer Assignee: Greg Studer
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
is related to SERVER-15892 Deadlock exceptions not caught in fin... Closed
Operating System: ALL
Participants:

 Description   

See: batch_executor.cpp::multiRemove()/multiUpdate(). The legacy write path is not necessarily affected as mongos no longer uses it.

If data has already been written these operations should not be aborted, and DeadlockExceptions can get thrown at any time during a multi-write.

Generally multi-writes are broadcast to all shards, but we should not rely on this behavior for correctness.



 Comments   
Comment by Greg Studer [ 04/Nov/14 ]

WAD - operation for deletes and updates now retries full operations except in the case of multi-updates, which cannot not throw stale exceptions on yield as in v2.6.

Generated at Thu Feb 08 03:39:11 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.